INTRODUCTIONDuring the period from birth to puberty, mammary gland

undergoes critical metamorphosis and obtains its future potentialto produce milk (1). Mammogenesis is allometric during thisinterval. At 3 months, the mammary gland starts a constant growthphase, much more rapidly than other parts of the body. At puberty,the proliferation of adipose tissue accompanies this rapid growthof the milk ducts. The first pre-pubertal stage is characterized byelongation and strengthening of the canals in the center of thefreshly tissue (2). This is extremely important to supply thesupport and space necessary to develop canals and ultimately, tosupport the frame on which the alveolar system is based. Onwardspuberty, the mammary gland follows an isometric growth (3). Atthe start of gestation, the growth of the mammary gland becomesallometric, because of an exponential increase in the numbers ofcells and canals. Following the elongation of the canals, thelobulo-alveolar system puts itself in place, progressivelysubstituting the adipose tissue that is in the process of regressing.In the last days of gestation, the epithelial cells increase in volumeand acquire specific structures to synthesize protein, to allowintense secretion. Many investigators have focused their studiesduring the prepubertal period (4, 5). During this critical stage, theincrease in mammary gland weight is correlated to the live weightgain (LWG) due to the adipose tissue increase (6, 7). On the otherhand, mammary parenchyma development is inverselyproportional to the live weight gain. Harrisson et al. (8) shown thatmammary parenchyma was permanently impaired by high rates ofLWG in the first year of life. This negative effect does not occurafter the puberty. Thereby, every alteration of mammogenesis

during the prepubertal using low/high feeding levels or hormonalmanipulations has an impact on the subsequent milk productionduring the first lactation.

In humans and rodents, evidence for role of estrogen inmammary gland development is considerable. The obligate roleof estradiol in mammogenesis has been underlined in ERαknocked out mice (9). In their experiment, authors demonstratedthat mammary glands from these mice were rudimentary withlimiting branching ducts and that this malformation couldn't bereversed by estradiol injection. Other studies showed that ERα isapproximately expressed in 15-30% of luminal epithelial cellsand not in other cells within the mammary gland (10, 11). Thissuggests that ERα is a key mediator of estradiol action in thenormal mammary gland. Berry et al. (12) have demonstrated thatovariectomy in prepubertal heifers caused a dramatic reductionin mammary gland development and in mammary epithelial cellproliferation. They have reported that ERα expression wasrestricted to mammary epithelial cells. Moreover, growthhormone (GH) injection can stimulate mammary epithelialproliferation, but no changes in the proportion of ERα-positivecells were found. Finally, their observations that mammarydevelopment was more severely affected in heifersovariectomized before 6 weeks of age implied that there is acritical period of ovarian stimulation during the first 2 month ofage. Ellis et al. (13) also concluded that parenchymal mammarydevelopment in the prepubertal lamb is largely unaffected byovariectomy, through roughly 3 months of age. They proposedovarian independence for prepubertal ovine mammarydevelopment and suggested that the endocrine control of ovinemammogenesis is distinctly different than in other species.

JOURNAL OF PHYSIOLOGY AND PHARMACOLOGY 2009, 60, Suppl 3, 127-133www.jpp.krakow.pl

F. DESSAUGE1, L. FINOT1, S. WIART1, J.M. AUBRY1, S.E. ELLIS2

EFFECTS OF OVARIECTOMY IN PREPUBERTAL GOATS

1INRA, Agrocampus Ouest, UMR1080 Dairy Production, Saint Gilles, France; 2Animal and Veterinary Sciences Department, Clemson University, Clemson, SC 29634, USA

The objectives of this study were to determine the effects of ovariectomy on mammary gland development in prepubertalgoats and to validate this model to study mammogenesis in young dairy ruminants. In this experiment, 3 months of agedgoats were ovariectomized (ovx) while shammed goats played as surgery controls (sham). Thereafter, sham and ovx goatswere slaughtered at 7 months of age to provide tissue for the assays. Results demonstrated that proliferation of mammaryof mammary epithelial cells was significantly lower in ovariectomized goats compared to control goats. In ovx animal,epithelium structures were completely overstretched and epithelial ducts were undeveloped with limited branching whereascontrol animals had classical complex arborescent units with multiple round ductules and limited stroma. Concerning ERα(estrogen receptor α), PR (progesterone receptor) and P450 (aromatase) expression, results showed number of ERα, PRand P450 positive cells was higher in shammed goats compared to ovariectomized goats. All this results suggested that goatmammogenesis and ovarian control are similar to prepubertal heifers and that young goats are a good model to studymammary gland development in ruminants. In conclusion, we demonstrated that ovariectomy of prepubertal goatsdecreased proliferation of mammary epithelial cells with a profound alteration of cell adhesion molecules.

K e y w o r d s : mammary gland, ovariectomy, prepubertal, goats, estrogen receptor

Mammary gland development in goats during theprepubertal period has not been extensively studied. The purposeof the present study was to increase the understanding ofbiological mechanisms underlying mammary growth anddevelopment as to elucidate the role of ovary during earlyprepubertal mammogenesis.

MATERIAL AND METHODSAnimals and tissue preparation

Mammary tissues used in this study were obtained from sixAlpine goats assigned to one of 2 treatments: shammed (sham)or ovariectomized (ovx). Surgery on animals was performed 3months after birth. Ovaries resection (ovx) concerned half goatsof the experiment whereas the other half was only opened andstitched (sham). Samples of mammary tissue were obtained atautopsy from animals at 7 months of age, under generalanaesthesia with subsequent euthanasia (Rompun i.v. 1 ml,Dolethal, i.v. 25 ml). The mammary glands from goats wereremoved within 20 minutes of slaughter. Freshly dissected tissuesegments were cut into small pieces (1 g) and frozen in liquidnitrogen for protein extraction or were fixed forImmunohistochemistry study. All procedures used in this studywere approved by French National Institute of AgriculturalResearch and by French Animal Care and Use Committee.

AntibodiesThe following antibodies were used for western blotting and

immunohistology: ERα (sc-787, Santa Cruz Biotechnology),P450 (sc-25270, Santa Cruz Biotechnology), PR (SM5025,Acris Antibodies GmbH), Cadherin-Catenin antibody samplerkit (#9961, Cell Signaling technology), PCNA (M0879,DakoCytomation), β-Actin (A5441, Sigma-Aldrich, France),anti-rabbit HRP-conjugated antibody (#7074, Cell Signalingtechnology), anti-mouse HRP-conjugated antibody (31450,Pierce, Perbio Science, Belgium) and goat anti-mouse FITC-conjugated antibody (F5387, Sigma-Aldrich, France).

Protein extractionTotal proteins were extracted from frozen mammary glands

tissues using the tissue protein extraction reagent T-PER (78510,Pierce, Perbio Science, Belgium). Mammary tissue sample (100mg) was ground in liquid nitrogen with a mortar and a pestle and 1ml of T-PER reagent was added to the powder. Afterhomogenization, the mixture was centrifuged at 10 000g for 5minutes at 4°C and the supernatant containing total proteins wasrecovered. The protein concentration was determined by the Lowrymethod (14) using the DC Protein Assay kit (500-0111, BioRad,Marnes-la-Coquette, France). Then, lysates were combined withsample buffer (50 mM Tris-HCl pH 6.8, 2% SDS, 0.1%bromophenol blue, 20% glycerol and 5% b-mercaptoethanol),boiled for 5 minutes at 95°C and resolved by SDS-PAGE.

Western blot analysisProteins (30 µg per lane) were separates on 10% SDS-

polyacrylamide gels, blotted to PVDF membranes (BioRad,Marnes-la-Coquette, France) and incubated with blockingsolution (5% dry skimmed milk dissolved in TBS-T (50 mM Tris-HCl pH: 8.6, 150 mM NaCl and 0.1% Tween) for 30 minutes. Aset of prestained molecular mass standards (SM0671, Fermentas,St Remy-les-chevreuses, France) was run in each gel.Membranes were incubated overnight at 4°C with the appropriate

dilution of the primary antibodies. Then, membranes werewashed with TBS-T before incubation with horseradishperoxidase-conjugated anti-mouse or anti-rabbit secondaryantibodies for 1 hour at room temperature. To detect theperoxidase activity, the enhanced chemiluminescence detectionsystem ECL (RPN2109, GE Healthcare Europe Gmbh, Saclay,France) was used. Membranes were then exposed to hyperfilm(Curix ortho HT-G films, Agfa, Mortsel, Belgium) and imagesgenerated were scanned at 16-bit/600dpi resolution with aCanoscan D1250 U2, saved as tiff files and calibrated to anoptical density scale. The integrated optical density of bands wasquantified using the ImageJ software. Each sample wasnormalized to β-actin content, a constitutively protein expressed.

Immunohistochemistry for proliferation and apoptosis assaysTo determine mammary epithelial cells (MEC) in

proliferation in the mammary gland, we carried outimmunohistochemical staining for PCNA (Proliferating CellNuclear Antigen). Indeed, proliferating mammary epithelialcells (MEC) were identified in mammary tissue as cellsexpressing the PCNA antigen as previously reported (15).Briefly, mammary gland tissues were fixed in PBS 4%paraformaldehyde for 24 h at 4°C, cryoprotected in 20% sucrosefor 48 h at 4°C, frozen in isopentane bath, cooled on dry ice, andstored at -80°C until use. Seven micrometer-thick cryosectionswere mounted onto Superfrost/Plus slides (Prolabo, Bondoufle,France) then quenched in PBS 3 % hydrogen peroxide (H2O2)and 10% methanol for 30 min. The sections were thoroughlywashed in PBS, permeabilized with PBS 1% SDS for 5 min,washed three times and pre-incubated in PBS 1% BSA for 1 h atroom temperature. The slides were incubated in the presence orabsence of the primary antibody for 1 night at 4°C. Afterwashing in PBS 1% BSA, samples were incubated with a secondantibody for 1 h at room temperature. Subsequently, themammary gland sections were counterstained for 3 min with 33µg/mL 4', 6-diamidino-2-phenylindole (DAPI, D9542, Sigma-Aldrich, France) then 3 min with propidium iodide at 333 µg/mL(P4864, Sigma-Aldrich, France).

Apoptosis induction in mammary cells was determined bythe terminal deoxynucleotidyltransferase (TdT)-mediated dUTPnick-end labeling (TUNEL) staining method based on DNAfragmentation detection. Seven micrometer-thick cryosectionsmounted onto 3-aminopropyltriethoxysilane (Sigma-Aldrich)treated slides, were unfrozen and incubated 30 min at 70°C in 10mM citrate sodium 0.1% triton solution. The slides were washedin PBS then incubated 30 min at 37°C in 200 ng/µL proteinaseK solution (V3021, Promega France, Charbonnieres, France)followed by incubation with the reagents of the DeadEndTMFluorometric TUNEL System (G3250, Promega France)according to the manufacturer's instructions. Mammary glandsections were counterstained with DAPI after TUNEL reaction.The slides were then mounted with Vectashield (Valbiotech,Paris, France), examined under fluorescence microscopy usingEclipse E400 Nikon microscope (Nikon France, Le Pallet,France) and pictures were captured by a Digital Still CameraDXM 1200 (Nikon). Height microscopic fields (magnification x200, area 0.14 mm2 per microscopic field) were examined foreach tissue sample. PCNA positive epithelial cells weredetermined thanks to IP (iodide propidium) staining whichoutlined the acini. Number of cells were detected and countedusing ImageJ software.

Immunohistochemistry for ERα, PR and P450 expressionCellular expression of ERα, PR and P450 were evaluated by

immunohistochemistry as described by Berry et al. (12). Briefly,

128

mammary gland tissues were fixed in PBS 4% paraformaldehydefor 24 h at 4°C and embedded in paraffin using standardprotocols. Five micrometers sections were mounted on positivelycharged slides, deparaffinized in xylene, rehydrated andfollowing rehydration, endogenous peroxidases of tissue sectionswere quenched in 3% H2O2. To unmask antigen sites, the slideswere microwaved for 3×5 min in 400 ml of 10 mM citrate buffer,pH 6.0. Slides were then cooled for 30 min, washed 3×2 min inPBS and blocked in 5% nonimmune serum for 30 min. Sectionswere incubated with 100 µl of primary antibodies overnight at4°C. Subsequently, slides were washed in PBS for 3×2 min anddetection of the primary antibody was performed using HRPimmunostaining kit (KP50L, Clinisciences, Montrouge, France).Slides were incubated with streptavidin-peroxidase (HRP)conjugate for 10 min, washed 3×2 min in PBS before visualizingthe antibody-HRP complex by incubation with diaminobenzidinefor 5 min. Finally, slides were dehydrated and mounted withPermount (Fisher Scientific, Pittsburgh, PA).

Statistical analysisData were analyzed by ANOVA using the mixed procedure of

SAS (SAS Institute, 1999) with repeated statement. The effect ofovariectomy was tested. Data were expressed as the mean ± SEM.

RESULTSHistology of mammary gland

Adipose tissue was widely represented compared tosecretory tissue (parenchyma) in ovariectomized goats (Fig. 1).Morphological analysis indicated that mammary parenchymaarea was very affected by ovariectomy. In ovx animals,epithelium structures were completely overstretched withundeveloped epithelial ducts with limited branching. In a sametime, control animals had classical complex arborescent unitswith multiple round ductules and limited stroma (adipocytesand connective tissue). Moreover, like bovine mammarystructures, there were three layers of cells: luminal cells, basalcells and embedded cells. We observed, in shammed animals,that these layers of epithelial cells were organized in ductules.Within the mammary gland from ovx animals, parenchymatissue is rudimentary and we observed a single layer ofepithelial cells.

Blood samples were collected weekly from birth to slaughterto measure concentration of estradiol. We found cyclicconcentration of circulating estrogen in control animals and theaverage of estrogen concentration was decreased 5-fold inovariectomized animals compared to control (data not shown).

129

Fig. 1. Representative images of eosinstaining of parenchymal tissue fromshammed (A) and ovariectomizedgoats (B). Left bar represents 20 µm.

Fig. 2. Expression of ERα in mammary epithelial cells. A)High magnification (bar represents 20 µm) of ductularstructures from shammed goat. Note expression of ERα inthe nucleus of epithelial cells only (brown staining). B) Highmagnification of ductular structures from ovariectomizedgoat. Note lower expression of ERα in the nucleus ofepithelial cells only (brown staining). C) Western blotanalysis of ERα in goat mammary gland extracts. Notedecreased expression of ERα in ovariectomized (OVX) goatscompared to shammed goats. Actin was used as control ofloading.

These results suggest that estradiol produced by ovaries has adirect effect on mammary gland development.

Effect of ovariectomy on ERα, PR and P450 aromataseexpression

It is well established that estrogen and progesterone areabsolutely essential for mammary epithelial proliferation anddifferentiation. In order to better understand the implication of

estrogen and progesterone during mammogenesis, we decided toinvestigate the distribution and expression pattern of estrogenreceptors (ERα) and progesterone receptors (PR).Immunolocalisation for the ERα revealed a positive staining inmammary epithelial cells of shammed animals and exclusivelylocalized in the nucleus of these cells (Fig. 2). Interestingly, verylow expression of ERα was found in ovx animals. Myoepithelialcells, adipocytes and cells from vascular system wereconsistently negative. Results obtained by Western Blot for ERα

130

Fig. 3. Expression of PR in mammary epithelial cells. A)Low magnification (bar represents 10 µm) of ductularstructures from shammed goat. Note expression of PR in thenucleus of epithelial cells only (brown staining). B) Lowmagnification of ductular structures from ovariectomizedgoat. Note lower expression of PR in the nucleus of epithelialcells only (brown staining). C) Western blot analysis of PR ingoat mammary gland extracts. Note decreased expression ofPR in ovariectomized (ovx) goats compared to shammedgoats. Actin was used as loading control.

Fig. 4. Expression of P450 aromatase in mammary epithelialcells. A) High magnification (bar represents 20 µm) ofductular structures from shammed goat. Note expression ofP450 in epithelial cells (brown staining). B) Highmagnification of ductular structures from ovariectomizedgoat. Note slightly lower expression of P450 in epithelialcells only (brown staining). C) Western blot analysis of P450in goat mammary gland extracts. Note slightly decreasedexpression of P450 in ovariectomized (ovx) goats comparedto shammed goats. Actin was used as control of loading.

in Fig. 2 confirmed that ERα (from mammary gland extracts)was higher expressed in shammed animals than in ovx animals.

Concerning progesterone receptor (PR) localization, wefound it localized in the nuclei of epithelial cells within themammary gland of shammed animals in a manner comparablewith that described for ER? expression (Fig. 3). Western blotresults indicated that PR expressed is stronger expressed inshammed animals compared to ovx animals.

Then we decided to identify the localization and theproduction of P450 aromatase within the mammary gland.Surprisingly, we observed a production of P450 aromatase in the

mammary epithelial cells at a high level both in shammed andovx animals (Fig. 4). P450 aromatase quantification by Westernblot showed more precisely that ovariectomy slightly affected itsexpression.

Effect of ovariectomy on epithelial cell proliferation andapoptosis

Proliferating cell nuclear antigen (PCNA) is highlyconserved auxiliary protein for DNA polymerase and is greatlyincreased in proliferation cells as compared with mitotically

131

Fig. 5. A) Proliferation of mammary epithelial cells asmeasured by positive PCNA staining into nucleus.Numbers of proliferating cells are expressed as apercentage of total epithelial cells. B) Apoptosis ofmammary epithelial cells as measured by TUNEL staininginto DNA. Numbers of apoptotic cells are expressed as apercentage of total epithelial cells. *, P

quiescent cells (16, 17). Wolf et al. (18) demonstrated a linearcorrelation between PCNA indices and S-phase fractions asdetermined by bromodeoxyuridine incorporation in regeneratingliver. Using PCNA staining, we observed that proliferation ofmammary epithelial cells (Fig. 5A) was decreased 7-fold in ovxgoats compared to control goats (7.22 vs. 1.49%; P

activation of numerous transcription factors and thetransactivation of target genes including that which encodeimportant regulators of growth, survival and differentiation. Wehypothesized that β-catenin-Wnt pathway could be inactivatedduring mammary gland development in ovx animals.

In conclusion, all this results suggested that goatmammogenesis and ovarian control could be compared toprepubertal heifers and that young goats are really good modelto study mammary gland development in ruminants. The modelis less expensive than heifers and animals manipulations aremore manageable.

Acknowledgements: Authors would like to thank MichelChorho and Eric Siroux for animal assistance and knowledgeabout goats rearing. We thank also Marion Boutinaud for criticalreading of the manuscript.

Conflict of interest statement: None declared.

REFERENCES1. Akers RM, Ellis SE, Berry SD. Ovarian and IGF-I axis

control of mammary development in prepubertal heifers.Domest Anim Endocrinol 2005; 29(2): 259-267.

2. Sinha YN, Tucker HA. Mammary development and pituitaryprolactin level of heifers from birth through puberty andduring the estrous cycle. J Dairy Sci 1969; 52(4): 507-512.

3. Capuco AV, Ellis S, Wood DL, Akers RM, Garrett W.Postnatal mammary ductal growth: three-dimensionalimaging of cell proliferation, effects of estrogen treatment,and expression of steroid receptors in prepubertal calves.Tissue Cell 2002; 34(3): 143-154.

4. Sejrsen K. Relationships between nutrition, puberty andmammary development in cattle. Proc Nutr Soc 1994; 53(1):103-111.

5. Sejrsen K, Purup S, Vestergaard M, Foldager J. High bodyweight gain and reduced bovine mammary growth:physiological basis and implications for milk yield potential.Domest Anim Endocrinol 2000; 19(2): 93-104.

6. Lammers BP, Heinrichs AJ, Kensinger RS. The effects ofaccelerated growth rates and estrogen implants inprepubertal Holstein heifers on estimates of mammarydevelopment and subsequent reproduction and milkproduction. J Dairy Sci 1999; 82(8): 1753-1764.

7. Sejrsen K, Huber JT, Tucker HA, Akers RM. Influence ofnutrition of mammary development in pre- and postpubertalheifers. J Dairy Sci 1982; 65(5): 793-800.

8. Harrison RD, Reynolds IP, Little W. A quantitative analysisof mammary glands of dairy heifers reared at different ratesof live weight gain. J Dairy Res 1983; 50(4): 405-412.

9. Bocchinfuso WP, Korach KS. Mammary gland developmentand tumorigenesis in estrogen receptor knockout mice. JMammary Gland Biol Neoplasia 1997; 2(4): 323-334.

10. Clarke RB, Howell A, Potten CS, Anderson E. Dissociationbetween steroid receptor expression and cell proliferation inthe human breast. Cancer Res 1997; 57(22): 4987-4991.

11. Speirs V, Skliris GP, Burdall SE, Carder PJ. Distinctexpression patterns of ER alpha and ER beta in normalhuman mammary gland. J Clin Pathol 2002; 55(5): 371-374.

12. Berry SD, Jobst PM, Ellis SE, Howard RD, Capuco AV, AkersRM. Mammary epithelial proliferation and estrogen receptoralpha expression in prepubertal heifers: effects of ovariectomyand growth hormone. J Dairy Sci 2003; 86(6): 2098-2105.

13. Ellis S, McFadden TB, Akers RM. Prepuberal ovinemammary development is unaffected by ovariectomy.Domest Anim Endocrinol 1998; 15(4): 217-225.

14. Lowry OH, Rosebrough NJ, Farr AL, Randall RL. Proteinmeasurement with the Folin phenol reagent. J Biol Chem1951; 193(1): 265-275.

15. Colitti M, Venturini E, Gabai G, Stradaioli G, Stefanon B.Apoptosis and expression of related proteins in mammarygland of heifers during early lactation. Vet Res Commun2003; 27(Suppl. 1): 225-227.

16. Celis JE, Bravo R, Larsen PM, Fey SJ. Cyclin: a nuclearprotein whose level correlates directly with the proliferativestate of normal as well as transformed cells. Leuk Res 1984;8(2): 143-157.

17. Fairman MP. DNA polymerase delta/PCNA: actions andinteractions. J Cell Sci 1990; 95(Pt 1): 1-4.

18. Wolf HK, Dittrich KL. Detection of proliferating cell nuclearantigen in diagnostic histopathology. J Histochem Cytochem1992; 40(9): 1269-1273.

19. Purup S, Sejrsen K, Akers RM. Effect of bovine GH andovariectomy on mammary tissue sensitivity to IGF-I inprepubertal heifers. J Endocrinol 1995; 144(1): 153-158.

20. Steinberg MS, McNutt PM. Cadherins and their connections:adhesion junctions have broader functions. Curr Opin CellBiol 1999; 11(5): 554-560.

21. Zwijsen RM, Wientjens E, Klompmaker R, van der SmanSJ, Bernards R, Michalides RJ. CDK-independent activationof estrogen receptor by cyclin D1. Cell 1997; 88(3): 405-415.

22. Simpson ER. Role of aromatase in sex steroid action. J MolEndocrinol 2000; 25(2): 149-156.

23. Barth AI, Nathke IS, Nelson WJ. Cadherins, catenins and APCprotein: interplay between cytoskeletal complexes andsignaling pathways. Curr Opin Cell Biol 1997; 9(5): 683-690.R e c e i v e d : November 5, 2008A c c e p t e d : April 15, 2009Author's address: Frederic Dessauge, INRA, Agrocampus

Ouest, UMR1080 Dairy Production, F-35590 Saint Gilles,France; e-mail: Frederic.Dessauge@rennes.inra.fr

133